The specific aims of this 5 year program-based collaborative research project are to: 1. evaluate and compare the effects of a relapse prevention/health education program and Therapeutic Community on completion rates, retention rates, and behavioral outcomes; 2) evaluate and compare the effects of a three-month relapse prevention-health education program and a six-month relapse program on prevention/health education on retention rates, completion rates, and HIV-related behaviors): 3. evaluate and compare the effects of a six-month Therapeutic Community and an one year Therapeutic Community (treatment as usual on completion rate, and behavioral outcomes: and 4. evaluate and compare the effectiveness of the programs outlined above in reducing AIDS-risk behavior in program participants. The study comprises two controlled clinical trials and an observational study. Approximately 600 clients applying to Spectrum House and Marathon, Inc. respectively (total of 1,200 participants) will be recruited and then randomized to one of the two relapse programs (3 or 6 months) for Spectrum clients and one of two Therapeutic Community intervention (6 to 12 months) for Marathon, Inc. clients. This research will 1) document pre-treatment behaviors and treatment history; 2) document program processes, retention, and completion rates: and 3) explore the relationships between program length, program type, and behavioral outcomes at 3 and 12 months after discharge. We hypothesize that 10 the relapse program will be more effective than the Therapeutic Community in reducing relapse rates and other behavioral outcomes because a) it acknowledges the chronicity of drug abuse and b) it provides specific skills needed to anticipate triggers to relapse and to deal effectively with them: 2) that short term treatment will have higher retention and completion rates compared to long term treatment: and 3) among those completing the programs, long term treatment will be more effective in terms of relapse rates. The research will also explore relationships between demographic and other variables (age,sex, race, HIV status, baseline psychological profile, socioeconomic status, previous drug treatment, legal history, and educational/vocational history) and retention rates, completion rates, and behaviors after treatment.